JP2001008045A - Color signal processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce an error due to nonlinearity by applying one-dimension nonlinearity conversion to an input color signal to obtain a 2nd color signal and conducting weight summing so that a ratio of the 2nd color signal increases as the strength in an obtained desired color signal direction as to the input color signal is stronger. SOLUTION: A strength coefficient arithmetic section 21 obtains a strength coefficient (t) in a desired color signal direction in an input CMY color space. A 1st conversion section 22 is provided with a plurality of one-dimensional lookup tables to compensate nonlinearity a color signal in a desired color signal direction. A signal adder section 23 receives both an input color signal and an output color signal of the 1st conversion section 22 and sums them with a weight in response to the strength coefficient (t). The color signal of CMY with each 10-bit width outputted from the signal adder 23 is fed to a 2nd conversion section 30, which divides each color into a high-order 5-bit and a low-order 5-bit, and they are given a three-dimensional lookup table(3D-LUT) 31 and an interpolation arithmetic section 32, in which the print color signal with each 10-bit width of the CMY is generated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color signal processing device for inputting a color signal representing a color separation image obtained by separating a color image into three colors and performing signal processing.

[0002]

2. Description of the Related Art In recent years, in the field of printing, for example, a digital color signal obtained by photoelectrically reading a document image recorded on a reversal film is inputted so that a desired color signal for printing can be obtained. Of digital signal processing. In this case, a signal for printing is obtained from three color signals of cyan (C), magenta (M), and yellow (Y) obtained by first reading a document image roughly (this is referred to as prescan). Signal processing in accordance with a predetermined signal processing algorithm for obtaining a color signal representing a halftone% of four colors obtained by adding black (K) to C, M, and Y. C, M, after the processing according to
An image based on the color signals of the four colors Y and K is displayed on a display screen such as a CRT display. The image displayed on this display screen is used for observation by the operator.

The signal processing algorithm includes HL / SH density setting for setting the density of a highlight (HL) portion and a shadow (SH) portion of an image, setting of a tone curve for determining the output density with respect to the input density, UCR (Und
er Color Removal) operation, a K-plate generation operation, a color correction operation for color correction, and the like, from C, M, and Y three-color signals.
This is an algorithm for performing complex non-linear signal processing for generating signals of four colors of Y and K. In this signal processing algorithm, various parameters are set, and the color signal obtained by the prescan includes this signal. Based on the processing algorithm, signal processing according to various set parameters is performed.

[0004] When an image is displayed on the display screen as described above, the operator adjusts various parameters of the signal processing algorithm while observing the image so that the color and the like of the image become desired. I do. When a desired image is obtained, an original image on the underlying reversal film is read at a high resolution in accordance with the operation of the operator (this is called a fine scan), and a color signal is obtained. The color signal is subjected to signal processing by a signal processing algorithm using the various parameters adjusted as described above. The signal obtained as a result is sent to an image output device, and an image is output to a film for plate making.

Here, the color signal obtained by the fine scan has an enormous data amount, and the signal processing is performed on the enormous amount of the color signal based on the calculation according to the above-described complicated signal processing algorithm. Since the signal processing takes an enormous amount of time if performed, usually, prior to fine scan, the three color signals of C, M, and Y are indexed based on the above signal processing algorithm. A look-up table (LUT), which is substantially equivalent to the above-described signal processing algorithm, for obtaining four color signals of C, M, Y, and K is created, and the color signal obtained by the fine scan refers to the LUT. By doing so, it is converted into an output color signal. Here, the LUT has C,
The three color signals of M and Y are output as color signals of four colors of C, M, Y and K using the three color signals as indices. Because it is a color, it is called a three-dimensional LUT (3D-LUT).

When creating this 3D-LUT, if the 3D-LUT is created as it is, for example, C,
When each of the M and Y color signals is represented by 10 bits, 3D-L having a huge number of grid points of about 10 ⁹
UT is required. For this reason, usually, a 3D-frame having a grid point of only upper bits (for example, 3 × 10 ⁴ grid points) is used.
Create an LUT and create its 3D-LU based on the lower bits
An operation method of performing an interpolation operation on the output of T is adopted (see Japanese Patent Publication No. 58-16180).

[0007]

In the above-described processing flow, the color signals obtained by the prescan are subjected to signal processing based on a signal processing algorithm, and an image based on the resulting color signals is converted to a CRT image. When displayed on a display screen such as a display and used for operator observation, the operator adjusts the values of various parameters of the signal processing algorithm so as to obtain a desired image regarding color and the like while observing the image. Is common. In this case, since these parameters do not directly represent the color and the like of the output image, the operator adjusts while predicting which parameter and how much the parameter will change the color and the like of the output image. As a result, it is often difficult to adjust well, and it often takes time to adjust.

In order to improve this, a point on the image displayed on the display screen is designated by a mouse or the like, and the C, M, Y, K halftones of the point (signal processing by the signal processing algorithm) are performed. Is performed on the same display screen, and the operator directly adjusts the displayed dot%, that is, directly adjusts the color and the like of the point on the image,
It has been proposed to automatically reflect the result of the adjustment of the dot% on the parameter value of the signal processing algorithm (see Japanese Patent Application Laid-Open No. 4-253474). By adopting this method, the operator can directly adjust the level (halftone%) of the output color signal, so that the adjustment is simplified and the work efficiency is improved.

However, since the signal processing algorithm is a complicated non-linear process as described above, the 3D-LUT created from the signal processing algorithm is also non-linear. Therefore, even if the operator directly adjusts the halftone percentage during the pre-scan, the halftone percentage obtained by converting the color signal obtained by the fine scan by the 3D-LUT and the interpolation arithmetic processing is used by the operator. Even if adjusted to the image having the desired color at the time of pre-scan, the image recorded on the plate making film differs from the desired color at the time of the pre-scan. There is a risk that it will.

Japanese Patent Application Laid-Open No. 8-138030 discloses a 3D
-A one-dimensional lookup table (1D-LUT) for each of the three colors R, G, and B is arranged before the LUT, and the 1D-LUT absorbs a non-linear component, thereby causing non-linearity in the interpolation operation. It has been proposed to reduce errors. If a nonlinear component can be absorbed in the 1D-LUT by using this method, an error caused by nonlinearity in the interpolation operation can be reduced, and the dot% obtained by the interpolation operation is adjusted by the operator during the prescan. Can be approached.

However, the method disclosed in this publication is based on R,
Since each 1D-LUT tries to uniformly absorb nonlinearity in each color axis direction of G and B,
Although the non-linearity is absorbed on each of the G and B color axes, for example, R, R,
The non-linearity is not eliminated in the color direction in which at least two colors of G and B are involved, but rather, the color in which two or more colors are involved as a result of uniformly correcting the axial directions of R, G and B. Also in the direction, the correction in the axial direction of R, G, and B influences, and the color direction may be further away from the desired characteristics or the nonlinearity may be increased. Therefore, the 1D is provided before the 3D-LUT. Although the concept of placing an LUT to absorb non-linearity and reducing errors due to non-linearity in the interpolation operation can be evaluated, the method according to this publication excluding the R, G, and B axes does not allow the interpolation operation to be performed. However, errors due to nonlinearity cannot be reduced.

The present invention has been made in view of the above circumstances, and has as its object to provide a color signal processing apparatus capable of reducing an error due to nonlinearity in an interpolation operation.

[0013]

A color signal processing apparatus according to the present invention for achieving the above object is a color signal processing apparatus which receives a color signal representing a color separation image obtained by separating a color image into three colors and performs signal processing. In the signal processing device, an intensity coefficient calculation unit that obtains an intensity coefficient representing an intensity of the input first color signal in a predetermined color direction in a color space;
A first conversion unit that obtains a second color signal by performing a one-dimensional non-linear conversion on the color signal of the first color signal and an intensity coefficient obtained by an intensity coefficient calculation unit by the first color signal and the second color signal. And a signal adding unit for obtaining a third color signal by performing weighted addition so that the ratio of the second color signal increases as the intensity of the first color signal in the predetermined color direction increases. The color direction processing unit, and the third
And a second conversion section for performing a three-dimensional conversion on the color signal to obtain a fourth color signal.

The fact that the color signal processing apparatus of the present invention includes the first conversion section before the second conversion section is described in Japanese Patent Application Laid-Open No. Hei 8-138030.
This is equivalent to providing a 1D-LUT before T, but in the color signal processing device of the present invention, the input first color signal
An intensity coefficient representing an intensity in a predetermined color direction in the color space, that is, a color direction in which non-linearity is to be compensated, is obtained, and a first conversion unit (1D-LUT) is determined according to the intensity coefficient.
Is weighted and added to the input first color signal and the output second color signal. As a result, according to the color signal processing device of the present invention, the nonlinearity is compensated only in the predetermined color direction and its vicinity, and the one-dimensional nonlinear conversion for the nonlinearity compensation in the color direction is performed in other color directions. Can be avoided. For example, in the case of a person image, the skin color is generally an important color. In this case, the image processing apparatus includes a first conversion unit that absorbs nonlinearity of the skin. By doing so, when the operator specifies a desired skin color in the pre-scan and performs the interpolation calculation processing in the second conversion unit in the fine scan, a skin color with a small error almost similar to the skin color specified by the operator is obtained. For color directions other than flesh color, the effect of compensating for the non-linearity of flesh color can be avoided.

Here, in the color signal processing apparatus of the present invention, the first color signal is a digital signal having a predetermined bit width for each color, and the color direction processing section has a predetermined bit width. In the case of performing an arithmetic operation, the second conversion unit normally inputs a signal obtained by collecting a predetermined upper bit portion of the predetermined bit width of the third color signal for three colors. A three-dimensional look-up table for obtaining an output corresponding to the signal; and an output of the three-dimensional look-up table based on a lower bit portion of the third color signal excluding the upper bit portion of the predetermined bit width. And an interpolation operation unit for performing an interpolation operation.

That is, the color direction processing section performs signal processing with a bit width (for example, each of C, M, and Y color signals is 10 bits) necessary and sufficient to compensate for nonlinearity in a predetermined color direction. , Due to the problem of the memory capacity of the second conversion unit, only the bit portion (for example, C,
A three-dimensional look-up table (3D-LUT) is provided which receives as input the upper 5 bits for each of the M and Y color signals, and a lower bit portion (for example, the lower 5 bits for each of the C, M and Y color signals). Is input to the interpolation operation unit,
The interpolation calculation unit is used for interpolation of the output of the 3D-LUT.

According to the present invention, the second conversion unit converts such a 3D-
The effect is remarkable in the case of a configuration in which the LUT and the interpolation calculation unit are combined, and the above-described predetermined color direction, that is, the color direction in which an error caused by nonlinearity caused by performing the interpolation calculation is reduced. With regard to the above, the dot% specified by the operator in the pre-scan and the dot% in the fine scan can be matched within a sufficiently small error range.

In the color signal processing apparatus of the present invention, typically, the first, second, and third color signals are obtained by separating a color image into three colors of cyan, magenta, and yellow. The second conversion unit inputs a third color signal representing three colors of cyan, magenta, and yellow and performs a three-dimensional conversion on the signal, which is a signal representing a color separation image of , Cyan, magenta, yellow, and black.

The color signal processing device of the present invention is suitable as a device for obtaining a color signal for printing.

Further, the color signal processing device of the present invention can be configured as a device for compensating for non-linearity in a plurality of different color directions. The thus configured color signal processing device of the present invention includes a plurality of color direction processing units that set mutually different color directions in the color space as the predetermined color direction, and includes the plurality of color direction processing units. Weighting processing for obtaining a fifth color signal by weighting and adding a plurality of third color signals for each color obtained in step (c) according to a plurality of intensity coefficients obtained in the plurality of color direction processing units. Wherein the second conversion unit performs three-dimensional conversion on the fifth color signal obtained by the weighting processing unit instead of the third color signal.

Thus, a plurality of color direction processing units are provided,
With the configuration in which the plurality of color signals obtained by the plurality of color direction processing units are weighted and added, the plurality of color directions processed by the plurality of color direction processing units are subjected to non-linearity compensation by the second conversion unit. It is possible to reduce an error caused by nonlinearity when performing an interpolation operation.

[0022]

Embodiments of the present invention will be described below.

FIG. 1 is a functional block diagram showing the configuration of the first embodiment of the color signal processing device of the present invention.

The color signal processing device 10 shown in FIG. 1 includes a color direction processing unit 20 and a second conversion unit 30. The color direction processing unit 20 includes an intensity coefficient calculation unit 21 and a first
It comprises a converter 22 and a signal adder 23.
The second arithmetic unit 30 includes a three-dimensional lookup table (3D-LUT) 31 and an interpolation arithmetic unit 32.

The color signal processing device 10 shown in FIG.
Three color signals of cyan (C), magenta (M), and yellow (Y) generated based on signals photoelectrically read by an image input device (not shown) are input. In the present embodiment, these three color signals are
It is a digital signal having a bit width of 10 bits for each color. C / C input to this color signal processing device
The M / Y color signal is input to each of the intensity coefficient calculation unit 21, the first conversion unit 22, and the signal addition unit 23 of the color direction processing unit 20.

In the intensity coefficient calculator 21, the input C /
An intensity coefficient t representing the intensity of the M / Y color signal in a predetermined color direction in the color space is obtained. Specific color direction,
A specific method of obtaining the intensity coefficient t for the color direction will be described later.

The first conversion unit 22 comprises a 1D-LUT for compensating for non-linearity in a predetermined color direction, or the 1D-LUT and another operation unit (to be described later).

Both the color signal input to the first conversion unit 22 and the color signal output from the first conversion unit 22 are input to the signal addition unit 23. t is also input, and the signal adder 23 weights and adds the color signal input to the first converter 22 and the color signal output from the first converter 22 with a weight corresponding to the intensity coefficient t.

The processing in the color direction processing unit 20 is C
The operation is performed with 10 bits being the bit width of each color signal of / M / Y.

C / M / Y output from the signal adder 23
, A 10-bit wide color signal for each of
It is input to the conversion unit 30. The C / M / Y color signal input to the second conversion unit 30 is divided into upper 5 bits and lower 5 bits for each color, and the upper 5 bits are the 3D-LUT3.
1 and the lower 5 bits are input to the interpolation calculator 32.

The 3D-LUT is a C / M / Y combination of the upper 5 bits (coordinates, grid points represented by the upper 5 bits on each axis of C / M / Y).
/ M / Y / K color signals, and 3D-L
When the upper 5 bits of each of the C / M / Y color signals are input to the UT 31, C / M / Y / K color signals corresponding to those input signals are read from the 3D-LUT 31. This 3D
The color signal read from the LUT 31 is
Is input to

The interpolation operation section 32 interpolates the color signals from the 3D-LUT 31 to obtain C / M / Y / K color signals at the points represented by the lower 5 bits of each color of C / M / Y. 1 for each color of C / M / Y / K
A 0-bit width color signal is output. This interpolation calculator 32
As an interpolation calculation algorithm in, for example, 3D
The output of the LUT 31 is obtained at four points for each output version of C / M / Y / K, and a weighting operation for weighting the four points with a weight based on the lower bit is performed. The algorithm disclosed in Japanese Patent No. 16180 can be adopted.

Here, the color direction processing unit 20 unique to the present invention is used.
Does not exist (in the case of only the second conversion unit 30 in the configuration shown in FIG. 1), the HL / SL density setting,
A signal processing algorithm including a combination of a tone curve, a color correction, and the like, and various parameters set for the signal processing algorithm are converted into a 3D-LUT 31. Here, the signal processing algorithm and the signal processing algorithm are converted. The various parameters set for the 3D-LUT 31 and the 1D-LUT provided in the first conversion unit 22 are converted separately. This conversion method will also be described later.

Next, various forms of the color direction processing section 20 will be described.

FIG. 2 is a diagram showing one embodiment of a gray direction processing unit, that is, a color direction processing unit for compensating for non-linearity in the gray direction. FIG. 3 is a diagram showing a one-dimensional lookup table (1D) shown in FIG. FIG. 3 is a diagram illustrating an example of a one-dimensional nonlinear conversion in (-LUT).

The gray direction processing section 201 includes a gray detecting section 211 (an example of the intensity coefficient calculating section 21 shown in FIG. 1),
It includes a 1D-LUT 221 (an example of the first conversion unit 22 shown in FIG. 1) and a signal addition unit 231. The signal adding unit 231 includes a subtractor 2311 and two multipliers 23
21 and 2331 and an adder 2341.

The gray detector 211 is one for the three colors C / M / Y, but the 1D-LUT 221 is C / M / Y.
And a total of three signal adders 231 are provided, one for each of the three colors C / M / Y.

[0038] In the gray detecting unit 211, when an input signal of the C / M / Y was C _in / M _in / Y _{in each, t g = Min (C in} , M in, Y in) / Max (C in , M
_in , Y _in ), the intensity coefficient t _{g in the} gray direction is obtained.

Here, Min (C _in , M _in , Y _in ) is
Max represents the minimum value of C _in , M _in , and Y _in , and Max
_{(C in, M in, Y} in) represents the maximum value of _{C in, M in, Y in} .

The 1D-LUT 221 is a look-up table having a relationship between an input and an output as shown in FIG. 3, for example, which differs for C / M / Y. C
/ M / Y 1D-LUT 221 conversion function (for example, FIG.
_{Are defined as} f _gc , f _gm , and f _gy , respectively. That is, when C _in , M _in , and Y _in are input to the 1D-LUT 221 of C / M / Y, respectively,
From the LUT, f _gc (C _in ), f _gm (M _in ), f
_gy (Y _in ) is output.

Each signal adder 231 corresponding to each of C / M / Y uses the intensity coefficient t _g in the gray direction obtained by the gray detector 211 as a weight, and calculates
M / Y color signals C _g , M _g , and Y _g are obtained.

[0042] _{C g = t g · f gc} (C in) + (1-t g) · C in M g = t g · f gM (M in) + (1-t g) · M in Y g = t _g · f _gY (Y _in ) + (1−t _g ) · Y _in To reduce the error due to nonlinearity in the interpolation operation _{in the} gray direction, the color direction processing unit 20 shown in FIG. Gray direction processing unit 201 shown in FIG.
Is adopted, and C _g / M _g / Y _g obtained by the gray direction processing unit 201 is input to the second conversion unit 30 shown in FIG.

FIG. 4 shows an example of a color direction processing unit, C / C.
FIG. 3 is a configuration diagram of a primary color direction processing unit that compensates for non-linearity in each primary color direction of M / Y.

The primary color direction processing unit 202 includes a C color direction processing unit for compensating for nonlinearity in the C direction, an M color direction processing unit for compensating for nonlinearity in the M direction, and a Y direction. The processing is divided into a Y color direction processing section for compensating for the non-linearity. Here, these are described together.

The primary color direction processing unit 202 is the same as the gray direction processing unit 201 shown in FIG.
2, a 1D-LUT 222 and a signal adder 232. The configuration of the signal addition unit 232 is
Is the same as the signal addition unit 231 of the gray color direction processing unit 201 shown in FIG.

In the primary color detecting section 212, three of C / M / Y
Assuming that the input color signal of the color is C _in / M _in / Y _in , the respective primary color intensity coefficients t _c , t _M , and t _Y for C / M / _Y are obtained as follows. That is, _{first, I p = (Max (C} in, M in, Y in) -Mid (C in, M in, Y in)) / Max (C in, M in, Y in) I c = (Max ( _{C in, M in, Y in} ) -Min (C in, M in, Y in)) / Max (C in, M in, Y in) where _{Mid (C in, M in,} Y in) is, C _in, M _in, Y
_Represents the central value of in. Here, Max (C
_in, when _{M in, Y in) = C} in, is found _{t c = I p · I c} t M = t Y = 0, Max (C in, M in, when the Y _in) = M _in, t _M = I _p · I _c t _c = t _Y = 0 is obtained, and when Max (C _in , M _in , Y _in ) = Y _in , t _Y = I _p · I _c t _c = t _M = 0 is required.

A total of three 1D-LUTs 222 are provided, one for each of C / M / Y. When color signals C _in , M _in , and Y _in are input to each of them, fD-LUT 222 has f
_c (C _in ), f _M (M _in ), and f _Y (Y _in ) are output.

The signal adder 232 performs the following weighted addition for each of C / M / Y, and performs C / M / Y
The Y color signals C _out , M _out , and Y _out are obtained.

[0049] _{C out = t c · f c} (C in) + (1-t c) · C in M out = t M · f M (M in) + (1-t M) · M in Y out = t _Y · f _Y (Y _in ) + (1−t _Y ) · Y _in FIG. 5 is an example of a color direction processing unit, which is a secondary color direction that compensates for non-linearity in each of R / G / B color directions. It is a block diagram of a processing part.

The secondary color direction processing unit 203 is divided into an R color direction processing unit, a G color direction processing unit, and a B color direction processing unit, as in the case of the primary color direction processing unit 202 shown in FIG. ,
Here, they are collectively described as a secondary color direction processing unit.

The secondary color direction processing unit 203 is similar to the gray color direction processing unit 201 shown in FIG. 2 and the primary color direction processing unit 202 shown in FIG.
It comprises an LUT 223 and a signal adder 233. The configuration of the signal addition unit 233 also includes the gray color direction processing unit 201 and the primary color direction processing unit 202 shown in FIGS.
Are the same as those of the signal adders 231 and 232.

In the secondary color detecting section 213, three of C / M / Y
Color input color signal is C_in/ M_in/ Y_inAnd the following
Thus, each secondary intensity for each of C / M / Y
Coefficient t_R, T_G, T_BIs required. That is, Min (C_in, M_in, Y_in) = C_inThen, t_R= (Mid (C_in, M_in, Y_in) -Min (C_in, M_in, Y_in)) / Max (C_in, M_in, Y_in) T_G= T_B= 0 Min (C_in, M_in, Y_in) = M_inThen, t_G= (Mid (C_in, M_in, Y_in) -Min (C_in, M_in, Y_in)) / Max (C_in, M_in, Y_in) T_R= T_B= 0 Min (C_in, M_in, Y_in) = Y_inThen, t_B= (Mid (C_in, M_in, Y_in) -Min (C_in, M_in, Y_in)) / Max (C_in, M_in, Y_in) T_R= T_G= 0 1D-LUT 222 is a two-dimensional color direction processing unit 203
Is the R color direction processing unit,
A total of two one-dimensional lookup tables,
In the one-dimensional lookup table, f
_RM(M_in), F_RY(Y_in) Is performed. Also,
When the two-dimensional color direction processing unit 203 is a G color direction processing unit
Corresponds to each of C / Y as 1D-LUT 222
Two 1D-LUTs are provided,
In the -LUT, f_GC(C_in), F _GY(Y_in)of
A conversion is performed. Furthermore, like this, its secondary color
When the direction processing unit 203 is the B color direction processing unit, 1D-L
As the UT 222, two 1s corresponding to the respective C / Ms
D-LUTs are provided and those 1D-LUTs
Is f_BC(C_in), F_BM(M_in) Conversion
Is

The two-dimensional signal adding section 233
When the color direction processing unit 203 is an R color direction processing unit, the following
By calculation, each color signal C of each of C / M / Y_R,
M_R, Y _RIs required.

[0054] _{C R = C in M R =} t R · f RM (M in) + (1-t R) · M in Y R = t R · f RY (Y in) + (1-t R) · Y _{in When the} secondary color direction processing section 203 is a G color direction processing section, the respective color signals C _G , M _G and Y _{G of} C / M / Y are obtained by the following calculation.

C _G = t _G · f _GC (C _in ) + (1−t _G ) · C _{in MG} = M _in Y _G = t _G · f _GY (Y _in ) + (1−t _G ) · Y _in Similarly, when the secondary color direction processing unit 203 is the B color direction processing unit, C / M /
Each of the respective color signals C _B of Y, M _B, the Y _B is determined.

[0056] _{C B = t B · f BC} (C in) + (1-t B) · C in M B = t B · f BM (M in) + (1-t B) · M in Y B = Y _in Figure 6, other than the above, for example, skin color, shows an example of a specific color direction processing unit for the particular color of the nonlinear compensation azure, etc., FIG. 7, the weighting factor calculation function for a particular color direction vector It is a figure showing an example of.

Here, in order to represent one or a plurality of specific colors in general, the n-th specific color is referred to as a specific color n. The specific color direction processing unit 204 includes the specific color detection unit 21
4. First conversion unit 224 for specific color, and signal addition unit 23
4 and the first conversion unit for specific color 224 includes:
It comprises a specific color component extraction unit 2214 and a 1D-LUT 2224.

Here, the vector in the direction of the specific color n in the color space in which the non-linearity is to be compensated by the 1D-LUT 2224 is defined as V _Pn (C _n , M _n , Y _n ).

The input color signals of C / M / Y are converted to C _in /
M _in / Y _in , the direction vector of the input color signal in the color space is V _i (C _in , M _in , Y _in ), and V _Pn and V _i
The angle theta _n constituting the ^{_{bets, θ n = cos -1 ((}} V Pn · V i) / | V Pn | · | V i |) obtained by. Here, (V _Pn · V _i ) represents the inner product of V _Pn and V _i .

The intensity coefficient t _Pn of the specific color n is obtained by t _Pn = 1−θ _n / θ _tn . Here, θ _tn is a constant of 0 <θ _tn <π / 2. The intensity coefficient t _Pn thus obtained is expressed as a function of θ _n as shown in FIG.

Here, specific numerical values of specific colors are exemplified as follows.

Example of skin color direction vector: V _P1 (120, 180, 22
0) Example of green direction vector: V _P2 (580, 360, 48
0) Example of sky blue direction vector: V _P3 (580, 360, 18
0) In the specific color component extraction unit 2214, the input color signal V
_{i (C in, M in,} Y in) direction component I _n of the specific color n of t _n = | obtained by calculating the cos [theta] _n | V _i.

The 1D-LUT 2224 is for each specific color n
And three for each of C / M / Y
These three 1D-LUTs have a specific color component extraction unit 2
Direction component I of specific color n obtained in 214_nIs entered
And f_pnC(I_n), F_{p nM}(I_n), F
_pnY(I_n).

The signal adder 234 performs the following operation to _obtain C _Pn , C / M / Y for each specific color n.
M _Pn and Y _Pn are determined.

[0065] _{C Pn = t Pn · f PnC} (I n) + (1-t Pn) · C in M Pn = t Pn · f PnM (I n) + (1-t Pn) · M in Y Pn = _{t Pn · f PnY (I n} ) + (1-t Pn) · Y in above, an example of a color direction processing unit on various color directions, of the color direction processing unit of each of these examples, Figure 1 The second shown in
A color direction processing unit that compensates for non-linearity in a color direction in which an error due to an interpolation operation in the interpolation operation unit 32 of the conversion unit 30 is a problem is adopted as the color direction processing unit 20 illustrated in FIG.

FIG. 8 is a block diagram showing the configuration of the second embodiment of the color signal processing device of the present invention.

The color signal processing device 100 includes a color direction processing unit group 200 composed of a plurality of color direction processing units, a weighting processing unit 400, and a second conversion unit 300.

The color direction processing unit group 200 is an example of the color direction processing unit according to the present invention. The gray direction processing unit 201 having the structure shown in FIG. 2 and the primary color C having the structure shown in FIG.
Direction processing unit 202C, primary color M direction processing unit 202M, 1
A secondary color Y direction processing unit 202Y, a secondary color R direction processing unit 203R, a secondary color G direction processing unit 203G having a configuration shown in FIG.
It comprises a secondary color B direction processing unit 203B, and a specific color one direction processing unit 204_1 and a specific color two direction processing unit 204_2 having the configuration shown in FIG.

The direction processing units 201, 202C,
202M, 202Y, 203R, 203G, 203B,
204_1 and 204_2 have a common input color signal V
_i (C _in , M _in , Y _in ) are input, and the respective intensity coefficients t _g , t _c , t _M , t _Y ,.
.., T _P2 are input to a coefficient normalization processing unit 410 constituting the weighting processing unit 400, and the respective C / M / Y color signals V _g , V _c , V _M obtained by the respective color direction processing units. ,
V _Y, ......, V _P2 (e.g. V _g is _{V g (C g, M g} , Y g) and a is the vector representation denoted _{.V c, V M, ......,} V P2
The same applies. ) Indicates the color direction processing units 201, 202C, 202
M, 202Y, 203R, 203G, 203B, 204
_1 and 204_2, and multipliers 401 and 402 included in the weighting processing unit 400, respectively.
C, 402M, 402Y, 403R, 403G, 403
B, 404_1 and 404_2.

In the coefficient normalization processing section 410, the input
All intensity factors t_g, T_c, T_M, T_Y, ……, t_P2In the sum of
Ratio u of each intensity coefficient to_g, U_c, U_M, U_Y, ……, u
_P2(E.g., with respect to the gray intensity factor, u_g= T_g/ (T
_g+ T_c+ T_M+ T_Y+ ... + t _P2). Other intensity factor t_c,
t_M, T_Y, ……, t_P2The same applies to ) Is required,
The corresponding multipliers 401, 402C, 402M,
402Y, 403R, 403G, 403B, 404_
1,404_2.

In each multiplier, each multiplier u shown in the figure is used._g・ V_g, U
_c・ V_c, U_M・ V_M, ……, u_P2・ V _P2(Specifically, the square
In the arithmetic unit 401, u_g・ C_g, U_g・ M_g, U_g・ Y_g
Represents three multiplications. The same applies to other multipliers.
Mr. ) Is required.

The result of the multiplication in all the multipliers is
Is input to the adder 420 which constitute the weighting processing unit 400 are added respectively with respect to C / M / Y, the color signal V _a
(C _a , M _a , Y _a ) is determined.

The adder 420 of the weighting section 400
Color signal V _a obtained in is input to the second conversion unit 300, the upper bit portion thereof that lower bit portion is input to the 3D-LUT 301 is inputted to the interpolation operation unit 302. The operation of the 3D-LUT 301 and the interpolation operation unit 302 of the second conversion unit 300 is the same as the operation of the 3D-LUT 31 and the interpolation operation unit 32 of the second conversion unit 30 described with reference to FIG. Description is omitted.

As shown in the second embodiment of FIG. 8, the present invention is also effective for reducing interpolation errors in a plurality of different color directions.

Next, how to calculate the 1D-LUT and the 3D-LUT will be described.

As described above, conventionally, usually, 3D-LU
In this case, since there is only T, each color signal corresponding to each grid point for C / M / Y is calculated according to a signal processing algorithm, and the calculation result is obtained.
An LUT can be determined. However, in the present embodiment, the non-linearity in the desired color direction is taken into the 1D-LUT from the signal processing algorithm (hereinafter, including various parameters unless otherwise specified), and the signal processing algorithm is realized as a whole. To be,
It is necessary to calculate both the 1D-LUT and the 3D-LUT. Therefore, here, 1D-LU is performed as follows.
T and 3D-LUT are calculated. (1) An output value corresponding to each input grid point is calculated according to a signal processing algorithm. Here, the signal processing algorithm is represented by F _IP , and the operation according to the signal processing algorithm is performed on the input color signals C _i , M _i , and Y _i by F _IP (C _IP
i , M _i , Y _i ) and the resulting C / M /
The Y / K color signal is simply represented by CMYK, and is represented as follows as a whole.

F _IP (C _i , M _i , Y _i ) | _CMYK (2) A direction component of interest is input, and a 1D-LUT is generated in correspondence with an output by a signal processing algorithm.
A 1D-LUT is generated for all necessary direction components.

For example, a 10-bit scale (0 to 102
3) In the case of gray, the C / M / Y input signal (C,
M, Y) in order (0, 0, 0), (1, 1, 1), (2, 2, 2),.
..., (1023, 1023, 1023) C / M of the output of the signal processing algorithm by changing
/ Y / K, 1D-LUT in which input C and output C are associated, 1D-L in which input M is associated with output M
-LUT and 1 in which input Y and output Y are associated
Create a D-LUT.

Regarding the primary colors C, M, and Y, for example, when explaining C, the input signals (C, M / Y) of C / M / Y
M, Y) in order (0, 0, 0), (1, 0, 0), (2, 0, 0),.
.., (1023, 0, 0)
M / Y / K is calculated, and a 1D-LUT in which the input C and the output C are associated is created. The same applies to M and Y.

Regarding the secondary colors R, G, and B, for example, when describing R, the C / M / Y input signals (C,
M, Y) in order (0, 0, 0), (0, 1, 1), (0, 2, 2),.
.., (0, 1023, 1023)
M / Y / K is calculated, and a 1D-LUT in which input M and output M are associated, and a 1D-LUT in which input Y and output Y are associated are created. The same applies to G and B.

Further, regarding the specific color n, the direction vector V _Pn (C _n , M _n , Y) of the characteristic color n in the C / M / Y color space
sequentially changing the length I _n of _n), the direction vector (I _n / each length I _n | V _Pn | C of) · V _Pn, M, Y components (I _{n
cosθ c, I n cosθ M,} I n cosθ Y; θ c, θ M,
theta _Y are each vector V _Pn and C, M, determine the angle formed) between the axes of Y, the set of components for each length I _n (I _n
cos [theta] _c, to calculate the _{I n cosθ M, I n cosθ} Y) is input to the signal processing algorithm C / M / Y / K. Then, to create a length I _n and the output of the C, M, Y 3 one 1D-LUT that associates respectively.

(3) The first conversion unit 2 of the color direction processing unit 20 shown in FIG. 1 (specifically, see FIGS. 2, 4, 5, and 6)
1D-L created as above in 2D-LUT
UT is applied, and each input signal V
_{i (C in, M in,} Y in) output color signal V _a of the color direction processing unit 20 sequentially inputs a _{(C a, M a, Y} a) Request.

[0083] The plurality of color directions when attempting to compensate for non-linearity, the output color signal V _a of the weighted addition unit 400 according to the configuration shown in FIG. _{8 (C a, M a,} Y a) Request.

(4) The color signal V thus obtained
_a (C _a , M _a , Y _a ) and C / M /
The 3D-LUT 3 shown in FIG.
1 or the 3D-LUT 301 shown in FIG. Various methods have been proposed for reconstructing a 3D-LUT of grid point input from the input / output relationship where the input is not a grid point (Japanese Patent Application Laid-Open Nos. 9-9080 and 10-70668).
JP-A-10-117294).

2 and 4 in the above embodiment.
The various color direction processing units shown in FIG. 6 to FIG. 6 are examples, and are not limited to those shown therein. For example, instead of the gray color direction processing unit shown in FIG. 2, gray may be regarded as one specific color and the configuration of the specific color direction processing unit shown in FIG. 6 may be applied, and various other modifications are possible.

FIG. 8 shows a configuration for reducing the interpolation errors in nine color directions. In the present invention, however, it is necessary to change the number of the color directions in which the interpolation errors are to be reduced and the number of the color directions. It is possible to provide a configuration including a necessary number of color direction processing units for compensating for the nonlinearity of the color direction.

[0087]

As described above, according to the present invention, in the method of obtaining the output value by interpolation calculation, the halftone output value of the screen display generated due to the nonlinear characteristic of the conversion and the actual signal processing. An error with the halftone output value can be reduced.

[Brief description of the drawings]

FIG. 1 is a functional block diagram illustrating a configuration of a first embodiment of a color signal processing device according to the present invention.

FIG. 2 is a diagram illustrating a gray direction processing unit.

FIG. 3 is a one-dimensional lookup table (1D) shown in FIG. 2;
FIG. 6 is a diagram illustrating an example of one-dimensional nonlinear conversion in (-LUT).

FIG. 4 is a configuration diagram of a primary color direction processing unit which is an example of a color direction processing unit.

FIG. 5 is a configuration diagram of a secondary color direction processing unit that is an example of a color direction processing unit.

FIG. 6 is a diagram illustrating an example of a specific color direction processing unit for compensating for non-linearity of a specific color such as flesh color or sky blue.

FIG. 7 is a diagram illustrating an example of a weight coefficient calculation function of a specific color direction vector.

FIG. 8 is a block diagram showing a configuration of a second embodiment of the color signal processing device of the present invention.

[Explanation of symbols]

Reference Signs List 10 color signal processing device 20 color direction processing unit 21 intensity coefficient calculation unit 22 first conversion unit 23 signal addition unit 30 second conversion unit 31 three-dimensional lookup table (3D-LUT) 32 interpolation calculation unit 200 color direction processing unit group Reference Signs List 201 Gray direction processing section 211 Gray detection section 221 1D-LUT 231 Signal addition section 2311 Subtractor 2321 Multiplier 2331 Multiplier 2341 Adder 202 Primary color direction processing section 202C Primary color C direction processing section 202M Primary color M direction Processing unit 202Y Primary color Y direction processing unit 212 Primary color detection unit 222 1D-LUT 232 Signal addition unit 203 Secondary color direction processing unit 203R Secondary color R direction processing unit 203G Secondary color G direction processing unit 203B B direction Processing unit 213 Secondary color detection unit 223 1D-LUT 233 Signal addition unit 204 Specific color direction processing unit 204_1 Specific color one-direction processing unit 204_2 Specific color two-direction processing unit 214 Specific color detection unit 224 First conversion unit for specific color 2214 Specific color component extraction unit 2224 1D-LUT 234 Signal addition unit 300 Second conversion unit 301 3D-LUT 302 interpolation calculation unit 400 weighting processing unit 401, 402C, 402M, 402Y, 403R, 4
03G, 403B, 404_1, 404_2 Multiplier 410 Coefficient normalization processing unit 420 Adder

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5B050 AA09 BA15 DA04 EA09 5B057 AA20 CA01 CA08 CA12 CB01 CB08 CB12 CC01 CD06 CD11 CH07 DB02 DB06 DB09 5C077 MP08 NN01 PP31 PP33 PQ11 PQ23 RR19 TT10 5C079 HB03 NA28

Claims (4)

[Claims] 1. A color signal processing device for inputting a color signal representing a color-separated image obtained by color-separating a color image into three colors and performing signal processing, wherein the input first color signal in a color space An intensity coefficient calculation unit for obtaining an intensity coefficient representing an intensity in a predetermined color direction, a first conversion unit for obtaining a second color signal by performing a one-dimensional nonlinear conversion on the input first color signal, The first color signal and the second color signal are divided into the second color signal and the second color signal in accordance with the intensity coefficient calculated by the intensity coefficient calculation unit, as the intensity of the first color signal in the predetermined color direction increases. A color direction processing unit including a signal addition unit that obtains a third color signal by weighting and adding so as to increase the ratio of the color signals; and a three-dimensional conversion into the third color signal obtained by the signal addition unit To obtain a fourth color signal. Color signal processing apparatus comprising the parts. 2. The method according to claim 1, wherein the first color signal is a digital signal having a predetermined bit width for each color, and the color direction processing unit performs an operation of the predetermined bit width. A three-dimensional look-up table, wherein the conversion unit inputs a signal obtained by collecting three predetermined color bits of a predetermined upper bit portion of the predetermined bit width of the third color signal and obtains an output corresponding to the signal; An interpolation operation unit that performs an interpolation operation on an output of the three-dimensional lookup table based on a lower bit portion of the third color signal, excluding the upper bit portion of the predetermined bit width. 2. The color signal processing device according to claim 1, wherein 3. The first, second, and third color signals are signals representing a color separation image obtained by separating a color image into three colors of cyan, magenta, and yellow, and the second color signal is a second color signal. The conversion unit inputs a third color signal representing three colors of cyan, magenta, and yellow and performs three-dimensional conversion, thereby obtaining four colors of cyan, magenta, yellow, and black for printing. 4. The color signal processing apparatus according to claim 1, wherein the color signal processing unit obtains four color signals. 4. A color space comprising a plurality of color direction processing units for setting mutually different color directions to the predetermined color direction, and a plurality of color direction processing units for each color determined by the plurality of color direction processing units. A weighting processing unit for obtaining a fifth color signal by weighting and adding the three color signals for each color in accordance with the plurality of intensity coefficients obtained by the plurality of color direction processing units; 2. The color signal processing apparatus according to claim 1, wherein the fifth color signal obtained by the weighting processing unit is subjected to three-dimensional conversion in place of the third color signal.